>Thanks for your reply Tim. I am currently involved in post-grad research in
>informatics (I have no background in biology) and being 'information
Oh, then consider moving into biological informatics -- A hot field.
>I have understood (or perhaps misunderstood) the process of biological
>evolution as a net increase in new & useful "code and data" within the
>genome over time.
It needn't be. Much of evolution probably results from relatively little
change in the "information content" of genomes. And there are not
very precise and useful metrics for genetic information -- at least
not within biological contexts. There may be some evidence to suggest
that vertebrates contain more information then invertebrates, but that's
not terribly firm. If we compare differences between vertebrates and
protists, I think a good case could be made. Eukaryotes vs. prokaryotes?
Probably, but I'd throw in the caveat that eukaryotes are actually
chimeric fusions, formed from mixing between two lines, one of which was
>Admittedly, the influencing factor has really been Lee Spetner's "Not
>By Chance" which did make a lot of sense. For example, how would a new
>functional feature arise without additional coding and data? (and I
>think this is in fact what Spetner means by 'information')
It is not clear what Spetner means by "information". He changes the
definition to suit his needs.
What is a "new" functional feature? Nobody studying evolution seriously
believes that complex and completely new features arises do novo in
lineages. It is generally thought that they arise as modifications from
previously existing features. A single point mutation can generate
a "new" feature. Antibiotic or disease resistance is one such feature.
(Spetner tries to rebut this example, but see previous posts by me
listed in the URLs posted below).
Can we get an enzyme to catalyze a reaction with a new substrate?
Yes. One simple way is duplication followed by relaxation of the
constraints for binding specificity -- allowing the enzyme to work
with a new substrate. This specificity can "tighten" to work primarily
with this other substrate. If we look at the many families of proteins
in cells today, we can see how this process of duplication and
divergence has lead to the creation and variation of enzymes.
Q: Spetner doesn't think duplication adds information. Is there
any informational metric where duplication doesn't add some
information? What if the duplication is imperfect and codes something
>As far as the evolution of the wing is concerned, how would scales evolve
>into feathers exactly? What is the actual mechanism? Is it simply natural
>selection acting on mutations? Surely new (or already existing but dormant)
>genetic code is needed for the transition.
"New" code in the sense that a particular sequence may have changed.
Not "new" in the sense that hundreds of bases suddenly popped into
existence in a couple steps. "Dormant" code gets obliterated by random
drift in a few million years and so I wouldn't rely on that as
a direct "pre-loaded" source of adaptive variation.
There are many ways of altering the timing, function and
position of expressed developmental pathways through mutation. For
starters these include: Deletion, insertion, single-base mutation,
inversion, transposition, and methylation. What was the specific
path from scales to feather? That's not possible to tell: 1) There
is no specific code "for" feathers. There's just a bunch a sequences
associated with the production of feathers -- some of these sequences
are probably used elsewhere in other pathways. 2)We don't know most
of the sequences associated with feather production. But we do know
that "new" features can arise from single mutations that affect
>Given environmental cues, the code is activated and morphological
>change is hence possible. Relevant "code and additional data" need
>to be present to 'convert' scales (input) into feathers (output) as
>initiated by environmental factors ('software flags'), given enough
Even this model assumes that new "additional data" becomes
incorporated into the genetic code in response to environmental
conditions. How is this all that different from standard models
of evolution -- with the exception that many models do not
assume environmental factors "initiate" specific genetic changes?
What if the "conversion" arises from random changes and becomes
selected by interaction with the environment?
The "added" information in these schemes come from interactions
between the organism and its environment. Even Spetner recognizes
a relationship between selection and information (though later to
ignore it in his book -- Think of how to quantify the information
gained by selection vs. the loss of information that Spetner claims
happens when a bacterium acquires resistance of streptomycin).
>I would appreciate any further comment, critique or clarification of
>any of the concepts here.
It's a variant of Lamarkism.
If you go back and read Spenter's book, see how he shifts his
definitions of "biological and genetic information". Basically,
he plays with different meanings of biological information in
such a way as to define any random change to the genome as "loss
Here's a reference to a letter from Spetner which Art Chadwick
posted to the evolution reflector at Calvin. This is where he
basically admits that he has no robust metric of biological
information. As you know, metrics are everything in this
line of work.
See also letters from August and September of this year
(look for Spetner in the titles), which discuss some of the
points you've presented. They're similar to what Art has
described in the past.
A review by Gert Korthof: